Gurujegan Murugesan (Leipzig) & Louise Raynaud (Göttingen)
Investigating the Anaphor-Agreement Effect: Icelandic ECM constructions and Agreement Switch in Kutchi-Gujarati
This presentation will feature two case-studies relating to the investigation of the Anaphor-
Agreement Effect, the descriptive generalisation made by Rizzi (1990) that anaphors may not
trigger regular co-varying phi-agreement.
Anaphor-Agreement Effect and Agreement Switch in Kutchi-Gujarati
Kutchi Gujarati exhibits a split agreement pattern just like many other Indo-Aryan languages.
In the imperfective, the agreement is with the subject and in the perfective, the agreement is
with the object (1a). It exhibits a particularly interesting instance of the Anaphor-Agreement
Effect (Rizzi 1990, Woolford 1999). In the perfective, reflexive objects seem unable to trigger
phi-covarying agreement on the verb. This results in agreement displacement/switch: perfective
verbs with reflexive objects agree with their subject (1b).Furthermore, when the agreement
switches to the subject and if the subject is a coordinated DP, then the agreement is with the
first conjunct (1c) (otherwise, it is resolved plural agreement).
(1) a. Mary John-ne ad-y-o
Mary John-ACC touch-PERF-M.SG
'Mary touched John.'
'John saw himself.'
c. [Mary ane John] pot-potha-ne jo-y-i/*a
Mary and John themselves-ACC see-PFV-F.SG/*PL
'Mary and John saw themselves.' (Patel 2014)
To account for these puzzles, we propose an alternative the analysis given in Patel (2014),
based on the one hand, on the aspectual split and the presence of a PerfP, and on the other hand,
an approach to conjunct agreement based on Van Koppen?s (2005, 2008) notion of equidistance.
Binding and Agreement in Icelandic ECM Constructions: From Nominative Reflexives to Pronouns
Nominative anaphors do not surface in Icelandic, not even in positions where nominative objects
may occur (possibly as a reflex of the AAE), giving rise to constructions in which pronouns can
behave as a reflexive. We observe such ECM constructions and try to provide an analysis as to
why the co-reference possibilities of the pronoun seem to correlate with the presence or absence
of covarying phi-agreement. We propose that (1a) and (1b) have different underlying structures
as a result of a restructuring operation.
(2) a. Konunumi fannst Þæri/j vera gáfaðar.
women.the.DAT found.3SG they.NOM be gifted.F.PL.NOM
'The womeni found them(selves)i/j to be gifted.'
b. Konunumi fundust Þær*i/j vera gáfaðar.
women.the.DAT found.3PL they.NOM be gifted.F.PL.NOM
'The womeni found them?i/j to be gifted.' (Taraldsen 1995: 317)